This pipeline is run as part of IRGen and has access to the IRGenModule.
Passes that run as part of this pipeline can query for the IRGenModule.
We will use it for the AllocStackHoisting pass. It wants to know if a type is of
non-fixed size.
To break the cyclic dependency between IRGen -> SILOptimizer -> IRGen that would
arise from the SILPassManager having to know about the createIRGENPASS()
function IRGen passes instead of exposing this function dynamically have to add
themselves to the pass manager.
Changes:
* Terminate all namespaces with the correct closing comment.
* Make sure argument names in comments match the corresponding parameter name.
* Remove redundant get() calls on smart pointers.
* Prefer using "override" or "final" instead of "virtual". Remove "virtual" where appropriate.
- TypeAliasDecl::getAliasType() is gone. Now, getDeclaredInterfaceType()
always returns the NameAliasType.
- NameAliasTypes now always desugar to the underlying type as an
interface type.
- The NameAliasType of a generic type alias no longer desugars to an
UnboundGenericType; call TypeAliasDecl::getUnboundGenericType() if you
want that.
- The "lazy mapTypeOutOfContext()" hack for deserialized TypeAliasDecls
is gone.
- The process of constructing a synthesized TypeAliasDecl is much simpler
now; instead of calling computeType(), setInterfaceType() and then
setting the recursive properties in the right order, just call
setUnderlyingType(), passing it either an interface type or a
contextual type.
In particular, many places weren't setting the recursive properties,
such as the ClangImporter and deserialization. This meant that queries
such as hasArchetype() or hasTypeParameter() would return incorrect
results on NameAliasTypes, which caused various subtle problems.
- Finally, add some more tests for generic typealiases, most of which
fail because they're still pretty broken.
The Linux build has a dependency on the libdispatch library,
which is needed by the various native libraries for sourcekitd.
On macOS, the dependency for libdispatch is satisfied directly through
the base OS, but on Linux no such dependency exists.
Modify this so that if the SourceKit library is built, and the
libdispatch library is already present, then we shell out to make
the libdispatch binary project when the SourceKit is built.
Issue: SR-1676
The sourcekitd-repl, sourcekitd-test and complete-test use symbols
from both the coverage and lto modules, specifically
`llvm::coverage::CoverageFilenamesSectionWriter::write` and
`llvm::lto::thinBackend`. Adding these to the list of dependencies
allows SourceKit to be built on Linux.
Issue: SR-3223
The errors won't be captured in the response, and it can be expensive to
do typo-correction when there are many errors, which is common in
indexing.
rdar://problem/28963058
Manually specifying range length is laborious and error-prone; this
commit adds an argument -end-pos=Line:Column to specify the end
position of a given range under test.
Like cursor-info, range info (""source.request.cursorinfo"") answers some
questions clients have for a code snippet under selection, for instance, the type of a selected
expression. This commit implements this new quest kind and provides two
simple information about the selected code: (1) the kind of the
snippet, currently limited to single-statement and expression; and (2)
the type of the selected expression. Gradually, we will enrich the
response to provide more insight into the selected code snippet.
Sorry -- I meant to include this with the previous commit. LLVM is
moving away from llvm::sys:TimeValue in favor of std::chrono. For consistency
with LLVM, I used the new llvm::sys::TimePoint type in most places.
There are a few places, like the SourceKit files modified here, where it
makes more sense to use std::chrono directly.
LLVM SVN r284681 replaced `LLVM_FUNCTION_NAME` with `__func__` as all supported
compilers support that keyword now. The holdout was MSVC, and swift requires
clang, so this is always supported. NFC.
Previously, document structure responses were created using XPC
dictionaries, but this forced deserialization of the large nested
dictionaries no matter what the access pattern was. The new format uses
our custom buffer machinery, and only deserializes dictionaries that are
actually needed.
The performance improvement is around 25% for an editor.open, and 35-40%
for an editor.replacetext request in the large files I've tested. The
performance improvements are in serialization (inside the service),
deserialization (client), and disposal of the response object (client).
rdar://problem/28789756
